Importing Libraries¶

In [1]:
import pandas as pd
import numpy as np
import missingno as msno
import matplotlib.pyplot as plt
import seaborn as sns
import plotly.express as px
import plotly.graph_objects as go
from plotly.subplots import make_subplots
import warnings
warnings.filterwarnings('ignore')

Imports from Scikit learn¶

In [2]:
from sklearn.preprocessing import StandardScaler
from sklearn.preprocessing import LabelEncoder

from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.naive_bayes import GaussianNB
from sklearn.neighbors import KNeighborsClassifier
from sklearn.svm import SVC
from sklearn.neural_network import MLPClassifier
from sklearn.ensemble import AdaBoostClassifier
from sklearn.ensemble import GradientBoostingClassifier
from sklearn.ensemble import ExtraTreesClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
from xgboost import XGBClassifier
from catboost import CatBoostClassifier
from sklearn import metrics
from sklearn.metrics import roc_curve
from sklearn.metrics import recall_score, confusion_matrix, precision_score, f1_score, accuracy_score, classification_report

Loading data¶

In [3]:
df = pd.read_csv('WA_Fn-UseC_-Telco-Customer-Churn.csv')
df.head()
Out[3]:
customerID gender SeniorCitizen Partner Dependents tenure PhoneService MultipleLines InternetService OnlineSecurity ... DeviceProtection TechSupport StreamingTV StreamingMovies Contract PaperlessBilling PaymentMethod MonthlyCharges TotalCharges Churn
0 7590-VHVEG Female 0 Yes No 1 No No phone service DSL No ... No No No No Month-to-month Yes Electronic check 29.85 29.85 No
1 5575-GNVDE Male 0 No No 34 Yes No DSL Yes ... Yes No No No One year No Mailed check 56.95 1889.5 No
2 3668-QPYBK Male 0 No No 2 Yes No DSL Yes ... No No No No Month-to-month Yes Mailed check 53.85 108.15 Yes
3 7795-CFOCW Male 0 No No 45 No No phone service DSL Yes ... Yes Yes No No One year No Bank transfer (automatic) 42.30 1840.75 No
4 9237-HQITU Female 0 No No 2 Yes No Fiber optic No ... No No No No Month-to-month Yes Electronic check 70.70 151.65 Yes

5 rows × 21 columns

In [4]:
df.shape
Out[4]:
(7043, 21)

The data set includes information about:¶

Customers who left within the last month – the column is called Churn

  • Services that each customer has signed up for – phone, multiple lines, internet, online security, online backup, device protection, tech support, and streaming TV and movies

  • Customer account information - how long they’ve been a customer, contract, payment method, paperless billing, monthly charges, and total charges

  • Demographic info about customers – gender, age range, and if they have partners and dependents

In [5]:
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 7043 entries, 0 to 7042
Data columns (total 21 columns):
 #   Column            Non-Null Count  Dtype  
---  ------            --------------  -----  
 0   customerID        7043 non-null   object 
 1   gender            7043 non-null   object 
 2   SeniorCitizen     7043 non-null   int64  
 3   Partner           7043 non-null   object 
 4   Dependents        7043 non-null   object 
 5   tenure            7043 non-null   int64  
 6   PhoneService      7043 non-null   object 
 7   MultipleLines     7043 non-null   object 
 8   InternetService   7043 non-null   object 
 9   OnlineSecurity    7043 non-null   object 
 10  OnlineBackup      7043 non-null   object 
 11  DeviceProtection  7043 non-null   object 
 12  TechSupport       7043 non-null   object 
 13  StreamingTV       7043 non-null   object 
 14  StreamingMovies   7043 non-null   object 
 15  Contract          7043 non-null   object 
 16  PaperlessBilling  7043 non-null   object 
 17  PaymentMethod     7043 non-null   object 
 18  MonthlyCharges    7043 non-null   float64
 19  TotalCharges      7043 non-null   object 
 20  Churn             7043 non-null   object 
dtypes: float64(1), int64(2), object(18)
memory usage: 1.1+ MB
In [6]:
df.columns.values
Out[6]:
array(['customerID', 'gender', 'SeniorCitizen', 'Partner', 'Dependents',
       'tenure', 'PhoneService', 'MultipleLines', 'InternetService',
       'OnlineSecurity', 'OnlineBackup', 'DeviceProtection',
       'TechSupport', 'StreamingTV', 'StreamingMovies', 'Contract',
       'PaperlessBilling', 'PaymentMethod', 'MonthlyCharges',
       'TotalCharges', 'Churn'], dtype=object)
In [7]:
df.dtypes
Out[7]:
customerID           object
gender               object
SeniorCitizen         int64
Partner              object
Dependents           object
tenure                int64
PhoneService         object
MultipleLines        object
InternetService      object
OnlineSecurity       object
OnlineBackup         object
DeviceProtection     object
TechSupport          object
StreamingTV          object
StreamingMovies      object
Contract             object
PaperlessBilling     object
PaymentMethod        object
MonthlyCharges      float64
TotalCharges         object
Churn                object
dtype: object

Visualizing missing values¶

From the visualisation we can observe that it has no peculiar pattern that stands out. In fact there is no missing data as obserbved below.

In [8]:
# Visualize missing values as a matrix
msno.matrix(df);
No description has been provided for this image

Manipulating and understanding data.¶

Dropping off Customer ID column

In [9]:
df = df.drop(['customerID'], axis = 1)
df.head()
Out[9]:
gender SeniorCitizen Partner Dependents tenure PhoneService MultipleLines InternetService OnlineSecurity OnlineBackup DeviceProtection TechSupport StreamingTV StreamingMovies Contract PaperlessBilling PaymentMethod MonthlyCharges TotalCharges Churn
0 Female 0 Yes No 1 No No phone service DSL No Yes No No No No Month-to-month Yes Electronic check 29.85 29.85 No
1 Male 0 No No 34 Yes No DSL Yes No Yes No No No One year No Mailed check 56.95 1889.5 No
2 Male 0 No No 2 Yes No DSL Yes Yes No No No No Month-to-month Yes Mailed check 53.85 108.15 Yes
3 Male 0 No No 45 No No phone service DSL Yes No Yes Yes No No One year No Bank transfer (automatic) 42.30 1840.75 No
4 Female 0 No No 2 Yes No Fiber optic No No No No No No Month-to-month Yes Electronic check 70.70 151.65 Yes

Analyzing in depth. Checking for null values in particular columns.

In [10]:
df['TotalCharges'] = pd.to_numeric(df.TotalCharges, errors='coerce')
df.isnull().sum()
Out[10]:
gender               0
SeniorCitizen        0
Partner              0
Dependents           0
tenure               0
PhoneService         0
MultipleLines        0
InternetService      0
OnlineSecurity       0
OnlineBackup         0
DeviceProtection     0
TechSupport          0
StreamingTV          0
StreamingMovies      0
Contract             0
PaperlessBilling     0
PaymentMethod        0
MonthlyCharges       0
TotalCharges        11
Churn                0
dtype: int64

Checking for TotalCharges Column.

In [11]:
df[np.isnan(df['TotalCharges'])]
Out[11]:
gender SeniorCitizen Partner Dependents tenure PhoneService MultipleLines InternetService OnlineSecurity OnlineBackup DeviceProtection TechSupport StreamingTV StreamingMovies Contract PaperlessBilling PaymentMethod MonthlyCharges TotalCharges Churn
488 Female 0 Yes Yes 0 No No phone service DSL Yes No Yes Yes Yes No Two year Yes Bank transfer (automatic) 52.55 NaN No
753 Male 0 No Yes 0 Yes No No No internet service No internet service No internet service No internet service No internet service No internet service Two year No Mailed check 20.25 NaN No
936 Female 0 Yes Yes 0 Yes No DSL Yes Yes Yes No Yes Yes Two year No Mailed check 80.85 NaN No
1082 Male 0 Yes Yes 0 Yes Yes No No internet service No internet service No internet service No internet service No internet service No internet service Two year No Mailed check 25.75 NaN No
1340 Female 0 Yes Yes 0 No No phone service DSL Yes Yes Yes Yes Yes No Two year No Credit card (automatic) 56.05 NaN No
3331 Male 0 Yes Yes 0 Yes No No No internet service No internet service No internet service No internet service No internet service No internet service Two year No Mailed check 19.85 NaN No
3826 Male 0 Yes Yes 0 Yes Yes No No internet service No internet service No internet service No internet service No internet service No internet service Two year No Mailed check 25.35 NaN No
4380 Female 0 Yes Yes 0 Yes No No No internet service No internet service No internet service No internet service No internet service No internet service Two year No Mailed check 20.00 NaN No
5218 Male 0 Yes Yes 0 Yes No No No internet service No internet service No internet service No internet service No internet service No internet service One year Yes Mailed check 19.70 NaN No
6670 Female 0 Yes Yes 0 Yes Yes DSL No Yes Yes Yes Yes No Two year No Mailed check 73.35 NaN No
6754 Male 0 No Yes 0 Yes Yes DSL Yes Yes No Yes No No Two year Yes Bank transfer (automatic) 61.90 NaN No

It can also be noted that the Tenure column is 0 for these entries even though the MonthlyCharges column is not empty.

Let's see if there are any other 0 values in the tenure column.

In [12]:
df[df['tenure'] == 0].index
Out[12]:
Index([488, 753, 936, 1082, 1340, 3331, 3826, 4380, 5218, 6670, 6754], dtype='int64')

Deleting the rows with missing values in Tenure columns since there are only 11 rows and deleting them will not affect the data.¶

In [13]:
df.drop(labels=df[df['tenure'] == 0].index, axis=0, inplace=True)
df[df['tenure'] == 0].index
Out[13]:
Index([], dtype='int64')

Filling the values of Totalcharges column with the mean values.

In [14]:
df.fillna(df["TotalCharges"].mean())
Out[14]:
gender SeniorCitizen Partner Dependents tenure PhoneService MultipleLines InternetService OnlineSecurity OnlineBackup DeviceProtection TechSupport StreamingTV StreamingMovies Contract PaperlessBilling PaymentMethod MonthlyCharges TotalCharges Churn
0 Female 0 Yes No 1 No No phone service DSL No Yes No No No No Month-to-month Yes Electronic check 29.85 29.85 No
1 Male 0 No No 34 Yes No DSL Yes No Yes No No No One year No Mailed check 56.95 1889.50 No
2 Male 0 No No 2 Yes No DSL Yes Yes No No No No Month-to-month Yes Mailed check 53.85 108.15 Yes
3 Male 0 No No 45 No No phone service DSL Yes No Yes Yes No No One year No Bank transfer (automatic) 42.30 1840.75 No
4 Female 0 No No 2 Yes No Fiber optic No No No No No No Month-to-month Yes Electronic check 70.70 151.65 Yes
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
7038 Male 0 Yes Yes 24 Yes Yes DSL Yes No Yes Yes Yes Yes One year Yes Mailed check 84.80 1990.50 No
7039 Female 0 Yes Yes 72 Yes Yes Fiber optic No Yes Yes No Yes Yes One year Yes Credit card (automatic) 103.20 7362.90 No
7040 Female 0 Yes Yes 11 No No phone service DSL Yes No No No No No Month-to-month Yes Electronic check 29.60 346.45 No
7041 Male 1 Yes No 4 Yes Yes Fiber optic No No No No No No Month-to-month Yes Mailed check 74.40 306.60 Yes
7042 Male 0 No No 66 Yes No Fiber optic Yes No Yes Yes Yes Yes Two year Yes Bank transfer (automatic) 105.65 6844.50 No

7032 rows × 20 columns

In [15]:
df.isnull().sum()
Out[15]:
gender              0
SeniorCitizen       0
Partner             0
Dependents          0
tenure              0
PhoneService        0
MultipleLines       0
InternetService     0
OnlineSecurity      0
OnlineBackup        0
DeviceProtection    0
TechSupport         0
StreamingTV         0
StreamingMovies     0
Contract            0
PaperlessBilling    0
PaymentMethod       0
MonthlyCharges      0
TotalCharges        0
Churn               0
dtype: int64

Labels for Senior Citizen column

In [16]:
df["SeniorCitizen"]= df["SeniorCitizen"].map({0: "No", 1: "Yes"})
df.head()
Out[16]:
gender SeniorCitizen Partner Dependents tenure PhoneService MultipleLines InternetService OnlineSecurity OnlineBackup DeviceProtection TechSupport StreamingTV StreamingMovies Contract PaperlessBilling PaymentMethod MonthlyCharges TotalCharges Churn
0 Female No Yes No 1 No No phone service DSL No Yes No No No No Month-to-month Yes Electronic check 29.85 29.85 No
1 Male No No No 34 Yes No DSL Yes No Yes No No No One year No Mailed check 56.95 1889.50 No
2 Male No No No 2 Yes No DSL Yes Yes No No No No Month-to-month Yes Mailed check 53.85 108.15 Yes
3 Male No No No 45 No No phone service DSL Yes No Yes Yes No No One year No Bank transfer (automatic) 42.30 1840.75 No
4 Female No No No 2 Yes No Fiber optic No No No No No No Month-to-month Yes Electronic check 70.70 151.65 Yes

Checking Categorical and numeric data.

In [17]:
df["InternetService"].describe(include=['object', 'bool'])
Out[17]:
count            7032
unique              3
top       Fiber optic
freq             3096
Name: InternetService, dtype: object
In [18]:
numerical_cols = ['tenure', 'MonthlyCharges', 'TotalCharges']
df[numerical_cols].describe()
Out[18]:
tenure MonthlyCharges TotalCharges
count 7032.000000 7032.000000 7032.000000
mean 32.421786 64.798208 2283.300441
std 24.545260 30.085974 2266.771362
min 1.000000 18.250000 18.800000
25% 9.000000 35.587500 401.450000
50% 29.000000 70.350000 1397.475000
75% 55.000000 89.862500 3794.737500
max 72.000000 118.750000 8684.800000

Visualizing the data¶

  • 26.6 % of customers switched to another firm.
  • Customers are 49.5 % female and 50.5 % male.
In [19]:
g_labels = ['Male', 'Female']
c_labels = ['No', 'Yes']
# Create subplots: use 'domain' type for Pie subplot
fig = make_subplots(rows=1, cols=2, specs=[[{'type':'domain'}, {'type':'domain'}]])
fig.add_trace(go.Pie(labels=g_labels, values=df['gender'].value_counts(), name="Gender"),
              1, 1)
fig.add_trace(go.Pie(labels=c_labels, values=df['Churn'].value_counts(), name="Churn"),
              1, 2)

# Use `hole` to create a donut-like pie chart
fig.update_traces(hole=.4, hoverinfo="label+percent+name", textfont_size=16)

fig.update_layout(
    title_text="Gender and Churn Distributions",
    # Add annotations in the center of the donut pies.
    annotations=[dict(text='Gender', x=0.19, y=0.5, font_size=20, showarrow=False),
                 dict(text='Churn', x=0.81, y=0.5, font_size=20, showarrow=False)])
fig.show()

No of users which churn or not by gender

In [20]:
df["Churn"][df["Churn"]=="No"].groupby(by=df["gender"]).count() # Did not churn
Out[20]:
gender
Female    2544
Male      2619
Name: Churn, dtype: int64
In [21]:
df["Churn"][df["Churn"]=="Yes"].groupby(by=df["gender"]).count() # churn
Out[21]:
gender
Female    939
Male      930
Name: Churn, dtype: int64

Churn distribution as per gender¶

There is negligible difference in customer percentage/ count who chnaged the service provider. Both genders behaved in similar fashion when it comes to migrating to another service provider/firm.

In [22]:
# Data
labels_outer = ['Churn: Yes', 'Churn: No']
values_outer = [1869, 5163]
colors_outer = ['#FF6F61', '#6BAED6']

labels_inner = ['Female (Yes)', 'Male (Yes)', 'Female (No)', 'Male (No)']
values_inner = [939, 930, 2544, 2619]
colors_inner = ['#E0BBE4', '#957DAD', '#AED9E0', '#7FC8A9']

# Create outer ring (Churn Yes/No)
fig = go.Figure()

fig.add_trace(go.Pie(
    labels=labels_outer,
    values=values_outer,
    hole=0.5,
    direction='clockwise',
    sort=False,
    marker=dict(colors=colors_outer, line=dict(color='#FFFFFF', width=2)),
    textinfo='label+percent',
    textfont_size=16,
    name='Churn',
    domain={'x': [0, 1], 'y': [0, 1]},
    hoverinfo='label+value+percent',
    showlegend=True
))

# Create inner ring (Gender breakdown)
fig.add_trace(go.Pie(
    labels=labels_inner,
    values=values_inner,
    hole=0.75,
    direction='clockwise',
    sort=False,
    marker=dict(colors=colors_inner, line=dict(color='#FFFFFF', width=2)),
    textinfo='label',
    textfont_size=14,
    name='Gender',
    domain={'x': [0, 1], 'y': [0, 1]},
    hoverinfo='label+value',
    showlegend=False
))

# Layout
fig.update_layout(
    title_text="Churn Distribution by Gender)",
    annotations=[dict(text='Churn', x=0.5, y=0.5, font_size=20, showarrow=False)],
    height=600,
    width=600
)

fig.show()

Contract distribution of customers¶

About 75% of customer with Month-to-Month Contract opted to move out as compared to 13% of customrs with One Year Contract and 3% with Two Year Contract

In [23]:
colors = ['#1f77b4', '#2ca02c', '#d62728']

fig = px.histogram(
    df, 
    x="Churn", 
    color="Contract", 
    barmode="group", 
    title="<b>Customer Contract Distribution</b>", 
    color_discrete_sequence = colors
)

fig.update_layout(
    width=700,
    height=500,
    bargap=0.1,
    # plot_bgcolor='#1e1e1e',  # dark background
    # paper_bgcolor='#1e1e1e',
    font=dict(color='black'),
    title_font=dict(size=20, color='black')
)

fig.show()

Payment method distribution¶

In [24]:
# Get labels and values
labels = df['PaymentMethod'].value_counts().index
values = df['PaymentMethod'].value_counts().values

# Colors: soft tones
colors = ['#66c2a5', '#fc8d62', '#8da0cb', '#e78ac3']

# Plot
plt.figure(figsize=(6.5, 6.5))
plt.pie(
    values,
    labels=labels,
    autopct='%1.1f%%',
    startangle=90,
    colors=colors,
    wedgeprops={'edgecolor': 'white', 'linewidth': 2},
    textprops={'fontsize': 13}
)

# Title and formatting
plt.title("Payment Method Distribution", fontsize=16, weight='bold')
plt.axis('equal')  # Equal aspect ratio ensures it's a circle
plt.tight_layout()
plt.show()
No description has been provided for this image

Payment method distribution w.r.t. churn¶

  • Major customers who moved out were having Electronic Check as Payment Method.
  • Customers who opted for Credit-Card automatic transfer or Bank Automatic Transfer and Mailed Check as Payment Method were less likely to move out.
In [25]:
dark_colors = ["#1f26b4", "#690e0e", "#093e09", "#ee7777"]

fig = px.histogram(
    df,
    x="Churn",
    color="PaymentMethod",
    title="<b>Customer Payment Method Distribution w.r.t. Churn</b>",
    color_discrete_sequence=dark_colors,
    barmode="stack" )

fig.update_layout(
    width=700,
    height=500,
    bargap=0.1,
    plot_bgcolor='white',
    paper_bgcolor='white',
    font=dict(color='black'),
    title_font=dict(size=18, color='black'),
    legend_title_text='Payment Method'
)

fig.show()

Churn distribution with internet services and gender¶

  • A lot of customers choose the Fiber optic service and it's also evident that the customers who use Fiber optic have high churn rate, this might suggest a dissatisfaction with this type of internet service.

  • Customers having DSL service are majority in number and have less churn rate compared to Fibre optic service.

In [26]:
df["InternetService"].unique()
Out[26]:
array(['DSL', 'Fiber optic', 'No'], dtype=object)
In [27]:
df[df["gender"]=="Male"][["InternetService", "Churn"]].value_counts()
Out[27]:
InternetService  Churn
DSL              No       992
Fiber optic      No       910
No               No       717
Fiber optic      Yes      633
DSL              Yes      240
No               Yes       57
Name: count, dtype: int64
In [28]:
df[df["gender"]=="Female"][["InternetService", "Churn"]].value_counts()
Out[28]:
InternetService  Churn
DSL              No       965
Fiber optic      No       889
No               No       690
Fiber optic      Yes      664
DSL              Yes      219
No               Yes       56
Name: count, dtype: int64
In [29]:
fig = go.Figure()

fig.add_trace(go.Bar(
  x = [['Churn:No', 'Churn:No', 'Churn:Yes', 'Churn:Yes'],
       ["Female", "Male", "Female", "Male"]],
  y = [965, 992, 219, 240],
  name = 'DSL',
))

fig.add_trace(go.Bar(
  x = [['Churn:No', 'Churn:No', 'Churn:Yes', 'Churn:Yes'],
       ["Female", "Male", "Female", "Male"]],
  y = [889, 910, 664, 633],
  name = 'Fiber optic',
))

fig.add_trace(go.Bar(
  x = [['Churn:No', 'Churn:No', 'Churn:Yes', 'Churn:Yes'],
       ["Female", "Male", "Female", "Male"]],
  y = [690, 717, 56, 57],
  name = 'No Internet',
))

fig.update_layout(title_text="<b>Churn Distribution w.r.t. Internet Service and Gender</b>")

fig.show()

Distribution w.r.t dependents

In [30]:
color_map = {"Yes": "#DEADDE", "No": "#2D1546"}
fig = px.histogram(df, x="Churn", color="Dependents", barmode="group", title="<b>Dependents distribution</b>", color_discrete_map=color_map)
fig.update_layout(width=700, height=500, bargap=0.1)
fig.show()

Churn distribution w.r.t. partners

In [31]:
color_map = {"Yes": "#672B2B", "No": "#2D2A63"}
fig = px.histogram(df, x="Churn", color="Partner", barmode="group", title="<b>Chrun distribution w.r.t. Partners</b>", color_discrete_map=color_map)
fig.update_layout(width=700, height=500, bargap=0.1)
fig.show()

Churn map w.r.t. senior citizen

  • It can be observed that the fraction of senior citizen is very less.
  • Most of the senior citizens churn.
In [32]:
color_map = {"Yes": "#1777A7", "No": "#A54367"}
fig = px.histogram(df, x="Churn", color="SeniorCitizen", title="<b>Chrun distribution w.r.t. Senior Citizen</b>", color_discrete_map=color_map)
fig.update_layout(width=700, height=500, bargap=0.1)
fig.show()

Churn w.r.t. online security

  • It has been observed that the customers churn in the absence of online security.
In [33]:
color_map = {"Yes": "#22A431", "No": "#DF5757"}
fig = px.histogram(df, x="Churn", color="OnlineSecurity", barmode="group", title="<b>Churn w.r.t Online Security</b>", color_discrete_map=color_map)
fig.update_layout(width=700, height=500, bargap=0.1)
fig.show()

Churn distribution w.r.t paperless billing

Customers with paperless billing are less likely to churn

In [34]:
color_map = {"Yes": '#1777A7', "No": '#A54367'}
fig = px.histogram(df, x="Churn", color="PaperlessBilling",  title="<b>Chrun distribution w.r.t. Paperless Billing</b>", color_discrete_map=color_map)
fig.update_layout(width=700, height=500, bargap=0.1)
fig.show()

Churn distribution with respect to Tech Support

Customers with no TechSupport are most likely to migrate to another service provide

In [35]:
fig = px.histogram(df, x="Churn", color="TechSupport",barmode="group",  title="<b>Chrun distribution w.r.t. TechSupport</b>")
fig.update_layout(width=700, height=500, bargap=0.1)
fig.show()

Churn distribution w.r.t. phone services

Very small fraction of customers don't have a phone service and out of that, 1/3rd Customers are more likely to churn.

In [36]:
color_map = {"Yes": '#00CC96', "No": '#B6E880'}
fig = px.histogram(df, x="Churn", color="PhoneService", title="<b>Chrun distribution w.r.t. Phone Service</b>", color_discrete_map=color_map)
fig.update_layout(width=700, height=500, bargap=0.1)
fig.show()

Distribution of monthly charges by churn

Customers with higher monthly charges are more likely to churn

In [37]:
sns.set_context("paper",font_scale=1.1)
ax = sns.kdeplot(df.MonthlyCharges[(df["Churn"] == 'No') ],
                color="Red", shade = True);
ax = sns.kdeplot(df.MonthlyCharges[(df["Churn"] == 'Yes') ],
                ax =ax, color="Blue", shade= True);
ax.legend(["Not Churn","Churn"],loc='upper right');
ax.set_ylabel('Density');
ax.set_xlabel('Monthly Charges');
ax.set_title('Distribution of monthly charges by churn');
No description has been provided for this image

Distribution of total charges by churn

In [38]:
ax = sns.kdeplot(df.TotalCharges[(df["Churn"] == 'No') ],
                color="Red", shade = True);
ax = sns.kdeplot(df.TotalCharges[(df["Churn"] == 'Yes') ],
                ax =ax, color="Blue", shade= True);
ax.legend(["Not Chu0rn","Churn"],loc='upper right');
ax.set_ylabel('Density');
ax.set_xlabel('Total Charges');
ax.set_title('Distribution of total charges by churn');
No description has been provided for this image

Tenure vs Churn

New customers are more likely to churn

In [39]:
fig = px.box(df, x='Churn', y = 'tenure')

# Update yaxis properties
fig.update_yaxes(title_text='Tenure (Months)', row=1, col=1)
# Update xaxis properties
fig.update_xaxes(title_text='Churn', row=1, col=1)

# Update size and title
fig.update_layout(autosize=True, width=750, height=600,
    title_font=dict(size=25, family='Courier'),
    title='<b>Tenure vs Churn</b>',
)

fig.show()
In [40]:
plt.figure(figsize=(25, 10))

corr = df.apply(lambda x: pd.factorize(x)[0]).corr()
mask = np.triu(np.ones_like(corr, dtype=bool))
ax = sns.heatmap(corr, mask=mask, xticklabels=corr.columns, yticklabels=corr.columns, annot=True, linewidths=.2, cmap='coolwarm', vmin=-1, vmax=1)
No description has been provided for this image

Data Preprocessing and Applying ML models¶

In [41]:
def object_to_int(dataframe_series):
    if dataframe_series.dtype=='object':
        dataframe_series = LabelEncoder().fit_transform(dataframe_series)
    return dataframe_series
In [42]:
df = df.apply(lambda x: object_to_int(x))
df.head()
Out[42]:
gender SeniorCitizen Partner Dependents tenure PhoneService MultipleLines InternetService OnlineSecurity OnlineBackup DeviceProtection TechSupport StreamingTV StreamingMovies Contract PaperlessBilling PaymentMethod MonthlyCharges TotalCharges Churn
0 0 0 1 0 1 0 1 0 0 2 0 0 0 0 0 1 2 29.85 29.85 0
1 1 0 0 0 34 1 0 0 2 0 2 0 0 0 1 0 3 56.95 1889.50 0
2 1 0 0 0 2 1 0 0 2 2 0 0 0 0 0 1 3 53.85 108.15 1
3 1 0 0 0 45 0 1 0 2 0 2 2 0 0 1 0 0 42.30 1840.75 0
4 0 0 0 0 2 1 0 1 0 0 0 0 0 0 0 1 2 70.70 151.65 1
In [43]:
plt.figure(figsize=(14,7))
df.corr()['Churn'].sort_values(ascending = False)
Out[43]:
Churn               1.000000
MonthlyCharges      0.192858
PaperlessBilling    0.191454
SeniorCitizen       0.150541
PaymentMethod       0.107852
MultipleLines       0.038043
PhoneService        0.011691
gender             -0.008545
StreamingTV        -0.036303
StreamingMovies    -0.038802
InternetService    -0.047097
Partner            -0.149982
Dependents         -0.163128
DeviceProtection   -0.177883
OnlineBackup       -0.195290
TotalCharges       -0.199484
TechSupport        -0.282232
OnlineSecurity     -0.289050
tenure             -0.354049
Contract           -0.396150
Name: Churn, dtype: float64
<Figure size 1400x700 with 0 Axes>
In [44]:
X = df.drop(columns = ['Churn'])
y = df['Churn'].values
In [45]:
X_train, X_test, y_train, y_test = train_test_split(X,y,test_size = 0.30, random_state = 40, stratify=y)
In [46]:
def distplot(feature, frame, color='r'):
    plt.figure(figsize=(8,3))
    plt.title("Distribution for {}".format(feature))
    ax = sns.distplot(frame[feature], color= color)

Since the numerical features are distributed over different value ranges, I will use standard scalar to scale them down to the same range.

In [47]:
num_cols = ["tenure", 'MonthlyCharges', 'TotalCharges']
for feat in num_cols: distplot(feat, df)
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image

Standardizing numeric attributes

In [48]:
df_std = pd.DataFrame(StandardScaler().fit_transform(df[num_cols].astype('float64')),
                       columns=num_cols)
for feat in numerical_cols: distplot(feat, df_std, color='g')
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image

Divide the columns into 3 categories, one ofor standardisation, one for label encoding and one for one hot encoding¶

In [49]:
cat_cols_ohe =['PaymentMethod', 'Contract', 'InternetService'] # those that need one-hot encoding
cat_cols_le = list(set(X_train.columns)- set(num_cols) - set(cat_cols_ohe)) #those that need label encoding
In [50]:
scaler= StandardScaler()

X_train[num_cols] = scaler.fit_transform(X_train[num_cols])
X_test[num_cols] = scaler.transform(X_test[num_cols])

Machine learning model evaluations and predictions¶

KNN¶

In [51]:
knn_model = KNeighborsClassifier(n_neighbors = 11) 
knn_model.fit(X_train,y_train)
predicted_y = knn_model.predict(X_test)
accuracy_knn = knn_model.score(X_test,y_test)
print("KNN accuracy:",accuracy_knn)

KNN accuracy: 0.776303317535545
In [52]:
print(classification_report(y_test, predicted_y))

              precision    recall  f1-score   support

           0       0.83      0.87      0.85      1549
           1       0.59      0.52      0.55       561

    accuracy                           0.78      2110
   macro avg       0.71      0.69      0.70      2110
weighted avg       0.77      0.78      0.77      2110

SVC¶

In [53]:
svc_model = SVC(random_state = 1)
svc_model.fit(X_train,y_train)
predict_y = svc_model.predict(X_test)
accuracy_svc = svc_model.score(X_test,y_test)
print("SVM accuracy is :",accuracy_svc)

SVM accuracy is : 0.8075829383886256
In [54]:
print(classification_report(y_test, predict_y))

              precision    recall  f1-score   support

           0       0.84      0.92      0.88      1549
           1       0.69      0.50      0.58       561

    accuracy                           0.81      2110
   macro avg       0.76      0.71      0.73      2110
weighted avg       0.80      0.81      0.80      2110

Random Forest¶

In [56]:
model_rf = RandomForestClassifier(n_estimators=500 , oob_score = True, n_jobs = -1,
                                  random_state =50, max_features = "sqrt",
                                  max_leaf_nodes = 30)
model_rf.fit(X_train, y_train)

# Make predictions
prediction_test = model_rf.predict(X_test)
print (metrics.accuracy_score(y_test, prediction_test))

0.8137440758293839
In [57]:
print(classification_report(y_test, prediction_test))

              precision    recall  f1-score   support

           0       0.84      0.92      0.88      1549
           1       0.71      0.51      0.59       561

    accuracy                           0.81      2110
   macro avg       0.77      0.72      0.74      2110
weighted avg       0.80      0.81      0.80      2110
In [60]:
plt.figure(figsize=(8,6))
sns.heatmap(confusion_matrix(y_test, prediction_test),
                annot=True,fmt = "d",linecolor="k",linewidths=3)
    
plt.title(" RANDOM FOREST CONFUSION MATRIX",fontsize=14)
plt.show()
No description has been provided for this image
In [61]:
y_rfpred_prob = model_rf.predict_proba(X_test)[:,1]
fpr_rf, tpr_rf, thresholds = roc_curve(y_test, y_rfpred_prob)
plt.plot([0, 1], [0, 1], 'k--' )
plt.plot(fpr_rf, tpr_rf, label='Random Forest',color = "r")
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.title('Random Forest ROC Curve',fontsize=16)
plt.show();
No description has been provided for this image

Logistic Regression¶

In [62]:
lr_model = LogisticRegression()
lr_model.fit(X_train,y_train)
accuracy_lr = lr_model.score(X_test,y_test)
print("Logistic Regression accuracy is :",accuracy_lr)

Logistic Regression accuracy is : 0.8090047393364929
In [63]:
lr_pred= lr_model.predict(X_test)
report = classification_report(y_test,lr_pred)
print(report)

              precision    recall  f1-score   support

           0       0.85      0.89      0.87      1549
           1       0.66      0.58      0.62       561

    accuracy                           0.81      2110
   macro avg       0.76      0.74      0.75      2110
weighted avg       0.80      0.81      0.80      2110
In [65]:
plt.figure(figsize=(8,6))
sns.heatmap(confusion_matrix(y_test, lr_pred),
                annot=True,fmt = "d",linecolor="k",linewidths=3)
    
plt.title("LOGISTIC REGRESSION CONFUSION MATRIX",fontsize=14)
plt.show()
No description has been provided for this image
In [66]:
y_pred_prob = lr_model.predict_proba(X_test)[:,1]
fpr, tpr, thresholds = roc_curve(y_test, y_pred_prob)
plt.plot([0, 1], [0, 1], 'k--' )
plt.plot(fpr, tpr, label='Logistic Regression',color = "r")
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.title('Logistic Regression ROC Curve',fontsize=16)
plt.show();
No description has been provided for this image

Decision tree classifier¶

In [67]:
dt_model = DecisionTreeClassifier()
dt_model.fit(X_train,y_train)
predictdt_y = dt_model.predict(X_test)
accuracy_dt = dt_model.score(X_test,y_test)
print("Decision Tree accuracy is :",accuracy_dt)

Decision Tree accuracy is : 0.7246445497630332
In [68]:
print(classification_report(y_test, predictdt_y))

              precision    recall  f1-score   support

           0       0.82      0.80      0.81      1549
           1       0.48      0.52      0.50       561

    accuracy                           0.72      2110
   macro avg       0.65      0.66      0.65      2110
weighted avg       0.73      0.72      0.73      2110

Ada Boost Classifier¶

In [69]:
a_model = AdaBoostClassifier()
a_model.fit(X_train,y_train)
a_preds = a_model.predict(X_test)
print("AdaBoost Classifier accuracy")
metrics.accuracy_score(y_test, a_preds)

AdaBoost Classifier accuracy
Out[69]:
0.8127962085308057
In [70]:
print(classification_report(y_test, a_preds))

              precision    recall  f1-score   support

           0       0.85      0.90      0.88      1549
           1       0.68      0.56      0.62       561

    accuracy                           0.81      2110
   macro avg       0.76      0.73      0.75      2110
weighted avg       0.81      0.81      0.81      2110
In [71]:
plt.figure(figsize=(8,6))
sns.heatmap(confusion_matrix(y_test, a_preds),
                annot=True,fmt = "d",linecolor="k",linewidths=3)
    
plt.title("AdaBoost Classifier Confusion Matrix",fontsize=14)
plt.show()
No description has been provided for this image

Gradient Boosting Classifier¶

In [72]:
gb = GradientBoostingClassifier()
gb.fit(X_train, y_train)
gb_pred = gb.predict(X_test)
print("Gradient Boosting Classifier", accuracy_score(y_test, gb_pred))

Gradient Boosting Classifier 0.8075829383886256
In [73]:
print(classification_report(y_test, gb_pred))

              precision    recall  f1-score   support

           0       0.85      0.90      0.87      1549
           1       0.67      0.55      0.60       561

    accuracy                           0.81      2110
   macro avg       0.76      0.73      0.74      2110
weighted avg       0.80      0.81      0.80      2110
In [75]:
plt.figure(figsize=(8,6))
sns.heatmap(confusion_matrix(y_test, gb_pred),
                annot=True,fmt = "d",linecolor="k",linewidths=3)
    
plt.title("Gradient Boosting Classifier Confusion Matrix",fontsize=14)
plt.show()
No description has been provided for this image

Voting Classifier¶

In [76]:
from sklearn.ensemble import VotingClassifier
clf1 = GradientBoostingClassifier()
clf2 = LogisticRegression()
clf3 = AdaBoostClassifier()
eclf1 = VotingClassifier(estimators=[('gbc', clf1), ('lr', clf2), ('abc', clf3)], voting='soft')
eclf1.fit(X_train, y_train)
predictions = eclf1.predict(X_test)
print("Final Accuracy Score ")
print(accuracy_score(y_test, predictions))

Final Accuracy Score 
0.8161137440758294
In [77]:
print(classification_report(y_test, predictions))

              precision    recall  f1-score   support

           0       0.85      0.90      0.88      1549
           1       0.68      0.57      0.62       561

    accuracy                           0.82      2110
   macro avg       0.77      0.74      0.75      2110
weighted avg       0.81      0.82      0.81      2110

From the confusion matrix we can see that: There are total 1400+149=1549 actual non-churn values and the algorithm predicts 1400 of them as non churn and 149 of them as churn. While there are 237+324=561 actual churn values and the algorithm predicts 237 of them as non churn values and 324 of them as churn values.

In [ ]:
cm = confusion_matrix(y_test, predictions)
labels = ['No Churn', 'Churn']

# Plot with pink-blue palette
plt.figure(figsize=(7, 5))
sns.heatmap(
    cm,
    annot=True,
    fmt='d',
    cmap=sns.color_palette(["#f3d7d9", "#6eaae2", "#669ab7", "#739ebd"], as_cmap=True),
    xticklabels=labels,
    yticklabels=labels,
    linewidths=2,
    linecolor='white',
    cbar=False,
    annot_kws={"size": 14, "weight": "bold", "color": "black"}
)

plt.title("Final Confusion Matrix", fontsize=16, weight='bold', color='#4B0082')
plt.xlabel("Predicted Label", fontsize=13)
plt.ylabel("True Label", fontsize=13)
plt.xticks(fontsize=12)
plt.yticks(fontsize=12)
plt.tight_layout()
plt.show()
No description has been provided for this image